PulseAugur / Brief
EN
LIVE 10:46:01

Brief

last 24h
[2/2] 224 sources

Multi-source AI news clustered, deduplicated, and scored 0–100 across authority, cluster strength, headline signal, and time decay.

  1. RESEARCH · arXiv cs.LG English(EN) · · [2 sources]

    When Do Autoregressive Sequence Models Forecast Physical Wavefields? A Controlled Study on Synthetic Seismograms

    Researchers have investigated the stability of autoregressive sequence models when forecasting long-horizon physical wavefields, such as seismograms. Their study, using a model called SeismoGPT on synthetic seismograms, found that multi-token prediction significantly stabilizes the forecasting process. Additional gains were observed with a horizon-embedding hybrid prediction head and a cross-horizon STFT-magnitude coherence loss, though performance critically depends on a specific context-ratio threshold. AI

    IMPACT Identifies key architectural choices for improving the stability of autoregressive models in long-horizon forecasting of physical signals.

  2. TOOL · arXiv cs.LG English(EN) ·

    Data-Driven Forecasting of three-Component Seismograms Using Transformer Architectures

    Researchers have developed SeismoGPT, a transformer-based model designed to forecast seismic waveforms. This model operates autoregressively in the time domain, continuing waveform data beyond observed seismic arrivals. SeismoGPT achieved high accuracy, with median normalized cross-correlation above 0.93, demonstrating its ability to maintain phase coherence and spectral energy distribution. The findings suggest that foundation models can be applied to physics-driven time-series forecasting, with potential uses in seismic warning and hazard mitigation, particularly for advanced gravitational-wave observatories. AI

    IMPACT Demonstrates potential for foundation models in physics-driven time-series forecasting, with applications in seismic warning systems.